A semantic encryption transmission method, device and equipment for perception information and communication information of a synaesthesia integrated platform
By adopting semantic encryption processing branch and semantic library recovery technology in the communication perception integrated platform, the problem of leakage risk between perceived information and communication information in coupled transmission is solved, and the secure transmission and privacy protection of information are realized.
Patent Information
- Application Number
- CN202410917905.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-10
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-07-10
AI Technical Summary
In the integrated communication and perception platform, the coupling of perceived information and communication information leads to the risk of leakage of sensitive perceived information and sensitive communication information, especially when unauthorized ports have the same computing power as legal ports, it is difficult to prevent information leakage.
The perceived semantic encryption processing branch and the communication semantic encryption processing branch are used to collect and preprocess information, select semantic keywords for encryption, and select a secure transmission path by calculating the average channel confidentiality rate, and restore and reconstruct using the preset semantic library to ensure that The integrity and privacy of information transmission.
Effectively prevent unauthorized visitors from stealing or tampering with sensitive perceived information and communication information, ensure the security and integrity of transmission, and improve the privacy protection of information.
Smart Images

Figure CN118473828B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of information security technology, and in particular to a semantic encryption transmission method, device and equipment for perception information and communication information of a synaesthesia integration platform. Background Art
[0002] The communication perception integrated routing system under wide coverage can break through the limitations of the traditional embedded sensor monitoring network on the geographical topology structure, and use the wave-particle duality of the WiFi frequency band to achieve seamless perception and communication coverage in the shadow range. However, the coupling of perception information and communication information under the communication perception integrated platform will cause physical layer security issues in terms of the leakage of sensitive perception information and sensitive communication information. When the unauthorized perception or unauthorized eavesdropping port has the same level of computing power as the target legitimate port, the unauthorized perception or unauthorized eavesdropping port is able to intercept and observe a small number of perception semantic keywords and communication semantic keywords, and use its perception and communication semantic library with the same computing power as the legitimate port to obtain a large amount of sensitive perception and communication information. Specifically, from the perspective of scene perception, when the semantic perception computing power of each port of the synaesthesia integrated platform is evenly distributed without difference, that is, the perception semantic library of these ports has the ability to restore the complete perception scene according to the perception semantic keywords, the legitimate perception port and the unauthorized perception port in the synaesthesia integrated platform can restore the perception scene through the perception semantic keywords transmitted in the network. At this time, the privacy protection of the perception semantic keywords is the key to preventing the leakage of sensitive perception content. From the perspective of communication information transmission, when the semantic communication computing power of each port of the synaesthesia integrated platform is evenly distributed without distinction, especially when unauthorized eavesdroppers in the communication network have a communication semantic library with the same computing power as the legitimate communication terminal, they all have the ability to completely restore the complete communication content based on the communication semantic keywords. At this time, the privacy protection of the communication semantic keywords is the key to preventing the leakage of sensitive communication information.
[0003] Therefore, how to prevent the leakage of sensitive perception information and sensitive communication information in the scenario where communication information and perception information are coupled is an issue that needs to be paid attention to in the current communication perception integrated platform in wide coverage scenarios. Summary of the invention
[0004] In view of this, the purpose of the present invention is to propose a semantic encryption transmission method, device and equipment for perception information and communication information of a synaesthesia integrated platform, aiming to solve the problem of how to improve the integrity and security of the transmission of perception information and communication information.
[0005] To achieve the above-mentioned object, the present invention provides a semantic encryption transmission method for perception information and communication information of a synaesthesia integrated platform, the method comprising a perception semantic encryption processing branch and a communication semantic encryption processing branch; wherein the perception semantic encryption processing branch comprises:
[0006] Collecting and preprocessing the perception information to obtain perception candidate keywords, and selecting perception semantic keywords from the perception candidate keywords;
[0007] Determine a sensitive perceptual semantic keyword among the perceptual semantic keywords, encrypt the sensitive perceptual semantic keyword, and select a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keyword by calculating an average channel confidentiality rate;
[0008] Recovering and reconstructing the received perception semantic keywords using a preset perception semantic library, and outputting the recovered perception information;
[0009] The communication semantic encryption processing branch includes:
[0010] Collecting and preprocessing communication information to obtain candidate communication keywords, and selecting communication semantic keywords from the candidate communication keywords;
[0011] Determine sensitive communication semantic keywords among the communication semantic keywords, encrypt the sensitive communication semantic keywords, and select a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating an average channel confidentiality rate;
[0012] The received communication semantic keywords are restored and reconstructed using a preset communication semantic library, and the obtained communication information is output.
[0013] Preferably, the collecting and preprocessing of the perception information to obtain the perception candidate keywords, and selecting the perception semantic keywords from the perception candidate keywords, includes:
[0014] Dividing the perception information into complete sentences and filtering each sentence to obtain the perception candidate keywords;
[0015] The perception candidate keywords are defined as a directed graph G including nodes and edges p =(V p , E p ), where V p represents the perception candidate keyword node set composed of the perception candidate keywords, E p It is a V p V p A set of , which represents the association value between the perceived candidate keyword nodes;
[0016] according to Calculate the score of each perceived candidate keyword node based on the directed graph, where is the damping coefficient, V pm Represents the point-to-point perception candidate keyword node V pi The node V pi 、V pj Represents the perceived candidate keyword node. For each perceived candidate keyword node V pi ,In(V pi ) represents the keyword set pointing to the node, Out(V pi ) represents the keyword set pointed to by the node, WS(V pj ) represents the score of the perceived candidate keyword at node j based on the directed graph, w p,ji 、w p,jm represents weight;
[0017] The weights of the perception candidate keyword nodes are iteratively propagated according to the scores until convergence, and the weights of each perception candidate keyword node are sorted, and the N perception candidate keywords with the highest weights are selected as the perception semantic keywords.
[0018] Preferably, the determining of the sensitive perceptual semantic keywords among the perceptual semantic keywords, encrypting the sensitive perceptual semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keywords by calculating the average channel confidentiality rate includes:
[0019] Determining sensitive perceptual semantic keywords among the perceptual semantic keywords according to the perceptual information sensitivity;
[0020] Using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive perception semantic keywords;
[0021] according to The average confidentiality rate of the channel is calculated, and the channel with the highest average confidentiality rate is used as a secure transmission path to transmit the perceptual semantic keyword, where R represents the average confidentiality rate of the channel, and the superscript + represents the positive part of the operation result. Indicates whether the receiver receives data in the nth time slot, SNR n,u and SNR n,s They are used to calculate the equivalent signal-to-noise ratio of the receiver to the sensor and the equivalent signal-to-noise ratio between the listener and the sensor in the nth time slot. lb(˙) represents the logarithm operation with base 2, which is equivalent to log2(˙).
[0022] Preferably, the using of a preset perceptual semantic library to restore and reconstruct the received perceptual semantic keywords and outputting the obtained restored perceptual information includes:
[0023] Using a dynamic multi-layer semantic perceptron model to semantically classify the perceptual semantic keywords, and obtain a perceptual semantic segmentation result;
[0024] Performing feature matching on each of the perceptual semantic keywords in the perceptual semantic segmentation result based on the perceptual semantic library, mapping the perceptual semantic keywords to the perceptual information space corresponding to the perceptual semantic library, and reconstructing through the features of the perceptual information space to obtain restored perceptual information;
[0025] A first mutual information between the restored perception information and the perception information is calculated, and a similarity between the restored perception information and the perception information is evaluated according to the first mutual information.
[0026] Preferably, the collecting and preprocessing of communication information to obtain communication candidate keywords, and selecting communication semantic keywords from the communication candidate keywords, includes:
[0027] Dividing the communication information into complete sentences and filtering each sentence to obtain the communication candidate keywords;
[0028] The communication candidate keywords are defined as a directed graph G including nodes and edges c =(V c , E c ), where V c represents the communication candidate keyword node set composed of the communication candidate keywords, E c It is a V c V c A set of,represents the association value between the communication candidate keyword nodes;
[0029] according to Calculate the score of each communication candidate keyword node based on the directed graph, where is the damping coefficient, V m Indicates the candidate keyword node V pointing to the communication ci The node V ci 、V cj Represents the communication candidate keyword node. For each communication candidate keyword node V ci ,In(V ci ) represents the keyword set pointing to the node, Out(V ci ) represents the keyword set pointed to by the node, w c,ji 、w c,jm represents weight, WS(Vcj ) represents the score of the communication candidate keyword at node j based on the directed graph;
[0030] The weight of each communication candidate keyword node is iteratively propagated according to the score until convergence, and the weight of each communication candidate keyword node is sorted, and the N communication candidate keywords with the highest weight are selected as the communication semantic keywords.
[0031] Preferably, the determining of sensitive communication semantic keywords among the communication semantic keywords, encrypting the sensitive communication semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating an average channel confidentiality rate includes:
[0032] Determining sensitive communication semantic keywords among the communication semantic keywords according to the sensitivity of the communication information;
[0033] Using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive communication semantic keywords;
[0034] according to The average confidentiality rate of the channels is calculated, and the channel with the highest average confidentiality rate is used as a secure transmission path to transmit the communication semantic keywords.
[0035] Preferably, the recovering and reconstructing the received communication semantic keywords by using a preset communication semantic library and outputting the obtained communication information comprises:
[0036] Using a dynamic multi-layer semantic perception model to semantically classify the communication semantic keywords, and obtain a communication semantic segmentation result;
[0037] Performing feature matching on each of the communication semantic keywords in the communication semantic segmentation result based on the communication semantic library, and mapping the communication semantic keywords to the communication information space corresponding to the communication semantic library, and reconstructing through the features of the communication information space to obtain restored communication information;
[0038] The second mutual information between the restored communication information and the communication information is calculated, and the similarity between the restored communication information and the communication information is evaluated according to the second mutual information.
[0039] To achieve the above-mentioned object, the present invention further provides a semantic encryption transmission device for perception information and communication information of a synaesthesia integrated platform, the device comprising a perception semantic encryption processing branch and a communication semantic encryption processing branch; wherein the perception semantic encryption processing branch comprises:
[0040] A first preprocessing unit, configured to collect and preprocess the perception information to obtain perception candidate keywords, and select perception semantic keywords from the perception candidate keywords;
[0041] A perceptual encryption unit, used for determining a sensitive perceptual semantic keyword among the perceptual semantic keywords, encrypting the sensitive perceptual semantic keyword, and selecting a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keyword by calculating an average channel confidentiality rate;
[0042] A perception reconstruction unit, used to restore and reconstruct the received perception semantic keywords using a preset perception semantic library, and output the obtained restored perception information;
[0043] The communication semantic encryption processing branch includes:
[0044] A second preprocessing unit is used to collect and preprocess communication information to obtain candidate communication keywords, and select communication semantic keywords from the candidate communication keywords;
[0045] A communication encryption unit, used for determining sensitive communication semantic keywords among the communication semantic keywords, encrypting the sensitive communication semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating an average channel confidentiality rate;
[0046] The communication reconstruction unit is used to recover and reconstruct the received communication semantic keywords using a preset communication semantic library, and output the obtained communication information.
[0047] In order to achieve the above-mentioned objectives, the present invention also proposes a semantic encryption transmission device for perception information and communication information of a synaesthesia integration platform, comprising a processor, a memory, and a computer program stored in the memory, wherein the computer program is executed by the processor to implement the steps of a semantic encryption transmission method for perception information and communication information of a synaesthesia integration platform as described in the above-mentioned embodiment.
[0048] In order to achieve the above-mentioned purpose, the present invention also proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of a semantic encryption transmission method of perception information and communication information of a synaesthesia integration platform as described in the above-mentioned embodiment.
[0049] Beneficial effects:
[0050] The above scheme pre-processes the perception information / communication information, selects keywords therein as perception semantic keywords / communication semantic keywords, determines sensitive perception semantic keywords / sensitive communication semantic keywords and encrypts them, transmits the perception semantic keywords / communication semantic keywords by calculating the channel with the highest average channel confidentiality rate, and restores and reconstructs the perception semantic keywords / communication semantic keywords at the receiving end using a preset semantic library, and outputs the restored perception information / communication information. This ensures that sensitive perception information and sensitive communication information are not stolen or tampered with by unauthorized visitors during the coupled transmission process, thereby ensuring the integrity and privacy of the content at both the perception information and communication information levels, and improving the security of transmission. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0052] Figure 1 A flowchart of a method for semantically encrypting and transmitting perception information and communication information of a synaesthesia integrated platform provided in one embodiment of the present invention.
[0053] Figure 2 A schematic diagram of the structure of a semantic encryption transmission device for perception information and communication information of a synaesthesia integration platform provided by an embodiment of the present invention.
[0054] The realization of the purpose, functional features and advantages of the invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0055] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0056] The present invention is described in detail below with reference to the embodiments.
[0057] In order to protect the security of perception semantic keywords, by combining encryption technology and secure transmission channels for information processing after the generation of perception information, perception information encryption strategy, transmission channel selection and receiver information recovery, it is possible to ensure that perception semantic keywords are not stolen or tampered with by unauthorized visitors during transmission, thereby ensuring the security and integrity of perception information and achieving encrypted transmission of perception semantic keywords. Similarly, in order to protect the security of communication semantic keywords, by combining encryption algorithms and strategies for the data sending end of the semantic communication system, communication information encryption strategy, encrypted transmission channel selection and receiver information recovery, it is possible to ensure the security and privacy of communication semantic keywords during transmission and achieve encrypted transmission of communication semantic keywords.
[0058] Reference Figure 1 The figure is a flow chart of a method for semantically encrypting and transmitting perception information and communication information of a synaesthesia integrated platform provided by an embodiment of the present invention.
[0059] In this embodiment, the method includes a perception semantic encryption processing branch and a communication semantic encryption processing branch; wherein the perception semantic encryption processing branch includes:
[0060] S11, collecting and preprocessing the perception information to obtain perception candidate keywords, and selecting perception semantic keywords from the perception candidate keywords.
[0061] Furthermore, the collecting and preprocessing of the perception information to obtain the perception candidate keywords, and selecting the perception semantic keywords from the perception candidate keywords, includes:
[0062] S11-1, dividing the perception information into complete sentences and filtering each sentence to obtain the perception candidate keywords;
[0063] S11-2, defining the perceived candidate keyword as a directed graph G including nodes and edges p =(V p , E p ), where V p represents the perception candidate keyword node set composed of the perception candidate keywords, E p It is a V p V p A set of , which represents the association value between the perceived candidate keyword nodes;
[0064] S11-3, according to Calculate the score of each perceived candidate keyword node based on the directed graph, where is the damping coefficient, V pm Represents the point-to-point perception candidate keyword node V pi The node V pi 、V pj Represents the perceived candidate keyword node. For each perceived candidate keyword node V pi ,In(V pi ) represents the keyword set pointing to the node, Out(V pi ) represents the keyword set pointed to by the node, w p,ji 、w p,jm represents the weight of the edge between adjacent nodes, WS(V pj ) represents the score of the perceived candidate keyword at node j based on the directed graph;
[0065] S11-4, iteratively propagating the weight of each perception candidate keyword node according to the score until convergence, and sorting the weight of each perception candidate keyword node, and selecting the N perception candidate keywords with the highest weight as the perception semantic keywords.
[0066] In this embodiment, the perception information is quantified into semantic keywords, that is, after the perception information is generated by the perception source including the sensor network or the data acquisition system, all the perception information is regarded as a perception word network, in which each perception keyword has a certain connection with other perception keywords in meaning. In order to effectively extract key information from this perception word network, it is necessary to analyze the entire network and read the words of all the perception information (using TextRank), and then extract the most representative and important perception semantic keywords based on the co-occurrence semantics between the words, so as to achieve the purpose of quantifying the semantic keywords of the perception information. Specifically:
[0067] The perceptual information generated by all perceptual sources is divided into complete sentences. In this process, the perceptual information is divided according to grammatical rules or punctuation marks, so that each sentence is a complete and independent semantic unit. In this way, the perceptual information is organized into a sentence sequence, denoted as T p =[S p1 , S p2 ,…,S pi ,…,S pn ], where each S pi Represents a complete sentence containing perceptual information. Each sentence is filtered and only perceptual keywords of specified parts of speech, such as nouns, verbs, and adjectives, are retained. The purpose is to filter out words with high perceptual information from the sentence and eliminate words that may appear in the perceptual information but have no practical meaning. The retained perceptual keywords will be used as perceptual candidate keywords for perceptual semantic keyword extraction.
[0068] Define the perceived candidate keywords as a directed graph G consisting of nodes and edges p =(V p , E p ), where V p Represents the node set composed of perceived candidate keywords, E p It is a V p V p The set of V represents the association value between the perceived candidate keyword nodes. For each perceived candidate keyword node V pi ,In(V pi ) represents the keyword set pointing to the node, Out(V pi ) represents the keyword set pointed to by the node. pi and V pj The association value between nodes is used as the weight w p,ji Add to Node V pi Based on the scores of the candidate keywords, the score ranking formula of the perceived candidate keyword nodes based on the graph is obtained, and the score of the perceived candidate keyword nodes based on the graph is calculated using the following formula:
[0069] ,
[0070] In the formula, is the damping coefficient, which represents the probability of jumping from a perception candidate keyword node to another random node in the graph, WS(V pj ) represents the score of the perceived candidate keyword at node j based on the directed graph.
[0071] According to the above score calculation formula, the weights of each perception candidate keyword node are iteratively propagated until convergence. During the iteration process, the weight value of each perception candidate keyword node is continuously updated to reflect the importance and influence of the perception candidate keyword in the entire network. In other words, from the arbitrary value (initial value) V assigned to each node (perception candidate keyword node) in the directed graph pj At the beginning, the calculation iterations are performed. After running the algorithm, each node has a value, which represents the "importance" of the node in the graph. In the process of weight propagation, the weights of the perception candidate keyword nodes are arranged in reverse order from large to small, and the N perception candidate keywords with the highest weights are obtained as the extracted perception semantic keywords. These perception semantic keywords carry rich perception information and meanings. In this way, the most valuable and meaningful information can be extracted from the massive amount of perception information.
[0072] S12, determining the sensitive perceptual semantic keywords among the perceptual semantic keywords, encrypting the sensitive perceptual semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keywords by calculating the average channel confidentiality rate.
[0073] Further, the determining of sensitive perceptual semantic keywords among the perceptual semantic keywords, encrypting the sensitive perceptual semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keywords by calculating an average channel confidentiality rate includes:
[0074] S12-1, determining sensitive perceptual semantic keywords among the perceptual semantic keywords according to the sensitivity of the perceptual information;
[0075] S12-2, using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive perception semantic keywords;
[0076] S12-3, according to The average confidentiality rate of the channel is calculated, and the channel with the highest average confidentiality rate is used as a secure transmission path to transmit the perceptual semantic keyword, where R represents the average confidentiality rate of the channel, and the superscript + represents the positive part of the operation result. Indicates whether the receiver receives data in the nth time slot, SNR n,u and SNR n,s They are used to calculate the equivalent signal-to-noise ratio of the receiver to the sensor and the equivalent signal-to-noise ratio between the listener and the sensor in the nth time slot. lb represents the logarithm operation with base 2, which is equivalent to log2(˙).
[0077] In this embodiment, some perceptual semantic keywords need to be encrypted before transmission, that is, according to the sensitivity of the perceptual information, that is, according to the importance of the information carried by each perceptual semantic keyword and the confidentiality requirements, so that the perceptual semantic keywords with more sensitive information and need to be kept confidential are divided into sensitive perceptual semantic keywords, and the others are divided into non-sensitive perceptual semantic keywords, so that the transmitted perceptual information has low sensitivity and resolution. Therefore, the selection of sensitive perceptual keywords is based on the established sensitivity classification optimization problem, and the optimal solution for the selection of sensitive perceptual keywords is obtained based on the optimal strategy. That is, by establishing a quantum approximate optimization algorithm, the quantum approximate optimization circuit model is used to consider the sensitivity factor of perceptual information to find the optimal encryption strategy. The transmission channel selection adopts the confidentiality rate pre-test transmission selection method. The channel with the highest channel confidentiality rate calculated in the transmission environment is used as a secure path to transmit all perceptual semantic keywords, ensuring that sensitive perceptual keywords will not be illegally obtained or eavesdropped during the transmission process. Specifically:
[0078] Consider the perceptual semantic keyword division as ap The keyword classification problem with sensitivity is represented by f p1 ,…,f pt F p , the keyword sensitivity target value is y p The goal of this classification problem is to select the k most sensitive p Keywords ( ). Define x p =(x p1 ,…,x pn ) {0,1} n Denotes the solution to the sensitive keyword selection problem, and defines Q p is an upper triangular matrix or n The symmetric matrix of n, Q p The expression of is calculated using the following formula:
[0079] ;
[0080] Among them, r(·,·) is the Pearson correlation coefficient. The support vector classifier is used to evaluate the sensitivity level. First, a support vector classifier is trained for each keyword sensitivity using the semantic keyword samples. Then, each trained classifier is used to classify each sensitive keyword into a sensitivity level to obtain the sensitivity results corresponding to the transmission-aware semantic keywords.
[0081] Set up a quadratic unconstrained binary optimization problem, and add a quadratic penalty term as the cost function of sensitive semantic keyword selection to make the actual number of selected keywords deviate from k p When , the value of the cost function increases. The resulting optimization problem based on sensitive perception semantic keyword selection is calculated using the following formula:
[0082] ; In the formula, represents the transposition operation of the solution to the sensitive keyword selection problem, Q p represents an upper triangular matrix or a symmetric matrix, x p represents the solution to the sensitive keyword selection problem, t p Indicates the number of keyword sensitivity levels, x pi represents the selected i-th sensitive keyword, k p Indicates the number of selected sensitive keywords.
[0083] The cost function is estimated by executing the quantum approximate optimization circuit model q times, and the Hamiltonian expectation of the final state of the system after each execution is measured to evaluate the pros and cons of each solution. Finally, the solution with the lowest cost is selected, which represents the best choice for the most sensitive semantic keyword.
[0084] Before transmitting the perceptual semantic keywords, it is necessary to select keyword training samples in batches for pre-transmission in the actual transmission scenario, and simulate and evaluate the possible monitoring situations before the actual transmission. The training samples include perceptual semantic keyword combinations for various perceptual scenarios. Based on the average channel confidentiality rate R result within one cycle obtained from the pre-transmission test, the channel path with the highest confidentiality is selected as the secure transmission path for data transmission of perceptual semantic keywords. The average confidentiality rate is calculated using the following formula: , where the Indicates whether the receiver receives data in the nth time slot, SNR n,u and SNR n,s They are used to calculate the equivalent signal-to-noise ratio of the receiver to the sensor and the equivalent signal-to-noise ratio between the listener and the sensor in the nth time slot.
[0085] S13, using a preset perceptual semantic library to restore and reconstruct the received perceptual semantic keywords, and outputting the obtained restored perceptual information.
[0086] Furthermore, the method of restoring and reconstructing the received perceptual semantic keywords by using a preset perceptual semantic library and outputting the obtained restored perceptual information includes:
[0087] S13-1, using a dynamic multi-layer semantic perception model to semantically classify the perceptual semantic keywords to obtain a perceptual semantic segmentation result;
[0088] S13-2, performing feature matching on each of the perceptual semantic keywords in the perceptual semantic segmentation result based on the perceptual semantic library, and mapping the perceptual semantic keywords to the perceptual information space corresponding to the perceptual semantic library, and reconstructing through the features of the perceptual information space to obtain restored perceptual information;
[0089] S13-3, calculating first mutual information between the restored perception information and the perception information, and evaluating similarity between the restored perception information and the perception information according to the first mutual information.
[0090] In this embodiment, all the perceptual semantic keywords collected by the receiving end are uploaded to the Tencent cloud server. The Tencent cloud server uses the received perceptual semantic keywords to analyze and match through the preset perceptual semantic library, and reconstructs and restores the content and meaning of the original perceptual information, and finally obtains complete perceptual information, realizing the security protection and privacy encryption transmission of perceptual information. The perceptual information recovered at the receiving end needs to be compared with the perceptual information generated by the perceptual source and the perceptual information recovery situation is evaluated by calculating the value of mutual information, and the encoder parameters are optimized by using the loss function to reduce the information transmission error. Specifically:
[0091] In order to more accurately identify and match perceptual semantic keywords in the subsequent perceptual information interaction process, Tencent Cloud needs to establish a perceptual semantic library based on both the sender and the receiver before the synaesthesia information interaction transmission. The perceptual semantic library content contains the previously transmitted perceptual information content. In addition, in each subsequent transmission and processing of perceptual information, Tencent Cloud needs to collect the newly received perceptual information and the corresponding perceptual semantic keywords, and add them to the perceptual semantic library for updating and supplementing, to ensure that the content in the perceptual semantic library can timely reflect the latest perceptual information content and semantic keyword changes, so that the perceptual information can be reconstructed and restored more accurately in the subsequent perceptual information interaction process.
[0092] Tencent Cloud has established a semantic information processing system based on deep learning to restore the received perceptual semantic keywords into specific and complete perceptual information using the perceptual semantic library. The system first uses a dynamic multi-layer semantic perceptron model for semantic classification and converts the previously existing perceptual semantic keywords x pi 0 Input to linear sequence conversion F pt (x pi 0 ), after obtaining the conversion result, each layer of semantic perceptrons and all base classifiers are trained, and after obtaining the perceptual semantic segmentation result, the corresponding perceptual semantic library is matched according to the features of each perceptual semantic keyword, and after matching, the perceptual semantic keywords are mapped to the perceptual information space of the corresponding perceptual semantic library, and the original perceptual information is reconstructed through the features of the perceptual information space to obtain the restored perceptual information for output.
[0093] At the receiving end, it is necessary to calculate the recovered perception information y p The sensory information x generated from the original sensory source pThe mutual information between the two can measure the correlation between the two information, evaluate the similarity and matching degree between the restored perceptual information and the original perceptual information, and objectively evaluate the effect of perceptual information recovery. Furthermore, in order to timely discover and correct possible errors and deviations and ensure that the quality of the restored perceptual information meets the expected requirements, the channel encoder C of the semantic information processing system can be optimized by maximizing the mutual information. α and semantic encoder S β , and use the loss function Loss(x p , y p ;α,β) train the neural network to obtain the encoder parameter settings α and β.
[0094] The communication semantic encryption processing branch includes:
[0095] S14, collecting and preprocessing communication information to obtain communication candidate keywords, and selecting communication semantic keywords from the communication candidate keywords.
[0096] Furthermore, the collecting and preprocessing of communication information to obtain communication candidate keywords, and selecting communication semantic keywords from the communication candidate keywords, includes:
[0097] S14-1, dividing the communication information into complete sentences and filtering each sentence to obtain the communication candidate keywords;
[0098] S14-2, defining the communication candidate keyword as a directed graph G including nodes and edges c =(V c , E c ), where V c represents the communication candidate keyword node set composed of the communication candidate keywords, E c It is a V c V c A set of,represents the association value between the communication candidate keyword nodes;
[0099] S14-3, according to Calculate the score of each communication candidate keyword node based on the directed graph, where is the damping coefficient, V m Indicates the candidate keyword node V pointing to the communication ci The node V ci 、V cj Represents the communication candidate keyword node. For each communication candidate keyword node V ci ,In(V ci ) represents the keyword set pointing to the node, Out(V ci) represents the keyword set pointed to by the node, w p,ji 、w p,jm represents the weight of the edge between adjacent nodes, WS(V cj ) represents the score of the communication candidate keyword at node j based on the directed graph;
[0100] S14-4, iteratively propagating the weight of each communication candidate keyword node according to the score until convergence, and sorting the weight of each communication candidate keyword node, and selecting the N communication candidate keywords with the highest weight as the communication semantic keywords.
[0101] In this embodiment, the communication information is quantified into semantic keywords, that is, after the communication information of the communication network is generated at the sending end, all the communication information is regarded as a communication word network, in which each communication keyword has a certain connection with other communication keywords in meaning. In order to effectively extract key information from this communication word network, it is necessary to analyze the entire network and read all the words containing communication information (using TextRank), and then extract the most representative and important communication semantic keywords based on the co-occurrence semantics between words, so as to achieve the purpose of quantifying the semantic keywords of communication information. Specifically:
[0102] All communication information generated by the sender is divided into complete sentences. In this process, the communication information will be divided according to grammatical rules or punctuation marks, so that each sentence is a complete and independent semantic unit. In this way, the communication information is represented as a sentence sequence, denoted by T c =[S c1 , S c2 ,…,S ci ,…,S cn ], where each S ci Represents a complete sentence containing communication semantic information. Each sentence is filtered and only communication keywords with specified parts of speech, such as nouns, verbs, and adjectives, are retained. Words with high communication information are selected, and words that may appear in the communication information but have no practical meaning, such as prepositions and articles, are eliminated. The retained communication keywords will be used as communication candidate keywords for communication semantic keyword extraction.
[0103] Define the communication candidate keywords as a directed graph G consisting of nodes and edges c =(V c , E c ), where V c Represents the node set composed of candidate keywords for communication, E c It is a V c V cThe collection represents the association value between the communication candidate keyword nodes. For each communication candidate keyword node V ci ,In(V ci ) represents the keyword set pointing to the node, Out(V ci ) represents the keyword set pointed to by the node. ci and V cj The association value between nodes is used as the weight w c,ji Add to Node V ci Based on the scores of the candidate keywords, the score ranking formula of the candidate keywords node based on the graph is obtained, and the score of the candidate keywords node based on the graph is calculated using the following formula:
[0104] ,
[0105] According to the above score calculation formula, the weights of each communication candidate keyword node are iteratively propagated until convergence. During the iteration process, the weight value of each communication candidate keyword node is continuously updated to reflect the importance and influence of the communication candidate keyword in the entire network. In the process of weight propagation, the weights of the communication candidate keyword nodes are arranged in reverse order from large to small, and the N communication candidate keywords with the highest weights are obtained as the extracted communication semantic keywords. These communication semantic keywords carry rich information and meanings, and can realize the compressed transmission of massive information.
[0106] S15, determining sensitive communication semantic keywords among the communication semantic keywords, encrypting the sensitive communication semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating the average channel confidentiality rate.
[0107] Further, the determining of sensitive communication semantic keywords among the communication semantic keywords, encrypting the sensitive communication semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating an average channel confidentiality rate includes:
[0108] S15-1, determining sensitive communication semantic keywords among the communication semantic keywords according to the sensitivity of the communication information;
[0109] S15-2, using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive communication semantic keywords;
[0110] S15-3, according to The average confidentiality rate of the channels is calculated, and the channel with the highest average confidentiality rate is used as a secure transmission path to transmit the communication semantic keywords.
[0111] In this embodiment, some communication semantic keywords need to be encrypted before transmission, that is, according to the sensitivity of the communication information, that is, according to the importance and confidentiality requirements of the information carried by each communication semantic keyword, the communication semantic keywords with more sensitive information and need to be kept confidential are classified as sensitive communication semantic keywords, while others are classified as non-sensitive communication semantic keywords, so that the transmitted communication information has low sensitivity and resolution. Therefore, the selection of sensitive communication keywords is based on the established sensitivity classification optimization problem, and the optimal solution for the selection of sensitive communication keywords is obtained based on the optimal strategy. That is, by establishing a quantum approximate optimization algorithm, the quantum approximate optimization circuit model is used to consider the sensitivity factor of communication information to find the optimal encryption strategy. The transmission channel selection adopts the confidentiality rate pre-test transmission selection method. The channel with the highest channel confidentiality rate calculated in the transmission environment is used as a secure path to transmit all communication semantic keywords, ensuring that sensitive communication keywords will not be illegally obtained or eavesdropped during the transmission process. Specifically:
[0112] Consider the communication semantic keywords as a c The keyword classification problem with sensitivity is represented by f c1 ,…,f ct F c , the keyword sensitivity target value is y c The goal of this classification problem is to select the k most sensitive c Keywords ( ). Define x c =(x c1 ,…,x cn ) {0,1} n Denotes the solution to the sensitive keyword selection problem, and defines Q c is an upper triangular matrix or n The symmetric matrix of n, Q c The expression of is calculated using the following formula:
[0113] ;
[0114] Among them, r(·,·) is the Pearson correlation coefficient. The support vector classifier is used to evaluate the sensitivity level. First, a support vector classifier is trained for each keyword sensitivity using the semantic keyword samples. Then, each trained classifier is used to classify each sensitive keyword into a sensitivity level to obtain the sensitivity results corresponding to the transmission communication semantic keywords.
[0115] Set up a quadratic unconstrained binary optimization problem, and add a quadratic penalty term as the cost function for sensitive communication semantic keyword selection so that the number of keywords actually selected deviates from k cWhen , the value of the cost function increases. The resulting optimization problem based on sensitive communication semantic keyword selection is calculated using the following formula: , where represents the transposition operation of the solution to the sensitive keyword selection problem, Q p represents an upper triangular matrix or a symmetric matrix, x c represents the solution to the sensitive keyword selection problem, t c Indicates the number of keyword sensitivity levels, x ci represents the selected i-th sensitive keyword, k c Indicates the number of selected sensitive keywords.
[0116] The cost function is estimated by executing the quantum approximate optimization circuit model q times, and the Hamiltonian expectation of the final state of the system after each execution is measured to evaluate the pros and cons of each solution. Finally, the solution with the lowest cost is selected, which represents the best choice for the communication semantic keyword with the highest sensitivity.
[0117] Before transmitting the communication semantic keywords, it is necessary to select communication semantic keyword training samples in batches for pre-transmission in the actual transmission scenario, and simulate and evaluate the possible monitoring situations before the actual transmission. The training samples include the communication semantic keyword combinations of various communication scenarios. Based on the average channel confidentiality rate R result within one cycle obtained from the pre-transmission test, the channel path with the highest confidentiality is selected as the secure transmission path for the communication semantic keyword data transmission. The average confidentiality rate is calculated using the following formula: , where the Indicates whether the receiving end receives data in the nth time slot, SNR n,u and SNR n,s They are used to calculate the equivalent signal-to-noise ratio of the receiver to the transmitter and the equivalent signal-to-noise ratio between the listener and the transmitter in the nth time slot.
[0118] S16, using a preset communication semantic library to recover and reconstruct the received communication semantic keywords, and outputting the obtained communication information.
[0119] Furthermore, the using of a preset communication semantic library to recover and reconstruct the received communication semantic keywords and outputting the obtained communication information includes:
[0120] S16-1, using a dynamic multi-layer semantic perception model to semantically classify the communication semantic keywords to obtain a communication semantic segmentation result;
[0121] S16-2, performing feature matching on each of the communication semantic keywords in the communication semantic segmentation result based on the communication semantic library, and mapping the communication semantic keywords to the communication information space corresponding to the communication semantic library, and reconstructing through the features of the communication information space to obtain restored communication information;
[0122] S16-3, calculating second mutual information between the restored communication information and the communication information, and evaluating similarity between the restored communication information and the communication information according to the second mutual information.
[0123] In this embodiment, all communication semantic keywords collected by the receiving end are uploaded to the Tencent cloud server. The Tencent cloud server uses the received communication semantic keywords to analyze and match through the preset communication semantic library, and reconstructs and restores the content and meaning of the original communication information, and finally obtains complete communication information, realizing the security protection and privacy encryption transmission of communication information. The communication information recovered at the receiving end needs to be compared with the communication information at the sending end and the value of mutual information is calculated to evaluate the recovery of the communication information, and the encoder parameters are optimized by using the loss function to reduce the information transmission error. Specifically:
[0124] In order to more accurately identify and match communication semantic keywords in the subsequent communication information interaction process, Tencent Cloud needs to establish a communication semantic library based on both the sender and the receiver before the synaesthesia information interaction transmission. The communication semantic library content contains the previously transmitted communication information content. And in each subsequent communication information transmission and processing process, Tencent Cloud needs to collect the newly received communication information and the corresponding communication semantic keywords, and add them to the communication semantic library for updating and supplementing, to ensure that the content in the communication semantic library can timely reflect the latest changes in the communication information content and semantic keywords, so as to more accurately reconstruct and restore the communication information in the subsequent communication information interaction process.
[0125] Tencent Cloud has established a semantic information processing system based on deep learning, which is used to restore the received communication semantic keywords into specific and complete communication information using the communication semantic library. The system first uses a dynamic multi-layer semantic perceptron model for semantic classification and converts the previously existing communication semantic keywords x ci 0 Input to linear sequence conversion Fc t (x ci 0 ), after obtaining the conversion result, each layer of semantic perception machine and all base classifiers are trained, and after obtaining the communication semantic segmentation result, the corresponding communication semantic library is matched according to the characteristics of each communication semantic keyword, and finally the communication semantic keyword is mapped to the communication information space of the corresponding communication semantic library, and the original communication information is reconstructed through the characteristics of the semantic space to obtain the restored communication information output.
[0126] At the receiving end, it is necessary to calculate the recovered communication information y c Communication information x generated by the sender c The mutual information between the two can measure the correlation between the two information, evaluate the similarity and matching degree between the restored communication information and the original communication information, and objectively evaluate the effect of communication information recovery. Furthermore, in order to timely discover and correct possible errors and deviations and ensure that the quality of the restored communication information meets the expected requirements, the channel encoder C of the semantic information processing system can be optimized by maximizing the mutual information. α and semantic encoder S β , and use the loss function Loss(x c , y c ; α,β) train the neural network to obtain the encoder parameter settings α and β.
[0127] Reference Figure 2 It is a schematic diagram of the structure of a semantic encryption transmission device for perception information and communication information of a synaesthesia integration platform provided by an embodiment of the present invention.
[0128] In this embodiment, the device 20 includes a perception semantic encryption processing branch and a communication semantic encryption processing branch; wherein the perception semantic encryption processing branch includes:
[0129] A first preprocessing unit 21 is used to collect and preprocess the perception information to obtain perception candidate keywords, and select perception semantic keywords from the perception candidate keywords;
[0130] A perceptual encryption unit 22, configured to determine a sensitive perceptual semantic keyword among the perceptual semantic keywords, encrypt the sensitive perceptual semantic keyword, and select a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keyword by calculating an average channel confidentiality rate;
[0131] A perception reconstruction unit 23, used to restore and reconstruct the received perception semantic keywords using a preset perception semantic library, and output the obtained restored perception information;
[0132] The communication semantic encryption processing branch includes:
[0133] A second preprocessing unit 24 is used to collect and preprocess communication information to obtain candidate communication keywords, and select communication semantic keywords from the candidate communication keywords;
[0134] A communication encryption unit 25, configured to determine sensitive communication semantic keywords among the communication semantic keywords, encrypt the sensitive communication semantic keywords, and select a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating an average channel confidentiality rate;
[0135] The communication reconstruction unit 26 is used to recover and reconstruct the received communication semantic keywords using a preset communication semantic library, and output the obtained communication information.
[0136] Each unit module of the device 20 can respectively execute the corresponding steps in the above method embodiment, so each unit module will not be described in detail here. Please refer to the description of the above corresponding steps for details.
[0137] The embodiment of the present invention also provides a semantic encryption transmission device for perception information and communication information of a synaesthesia integration platform, the device comprising the semantic encryption transmission device for perception information and communication information of the synaesthesia integration platform as described above, wherein the semantic encryption transmission device for perception information and communication information of the synaesthesia integration platform can adopt Figure 2 The structure of the embodiment can be executed accordingly. Figure 1 The technical solution of the method embodiment shown has similar implementation principles and technical effects. For details, please refer to the relevant records in the above embodiments, which will not be repeated here.
[0138] The device includes: a mobile phone, a digital camera, a tablet computer, or other devices with a camera function, or a device with an image processing function, or a device with an image display function. The device may include components such as a memory, a processor, an input unit, a display unit, and a power supply.
[0139] Among them, the memory can be used to store software programs and modules, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as an image playback function, etc.), etc.; the data storage area may store data created according to the use of the device, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory may also include a memory controller to provide the processor and the input unit with access to the memory.
[0140] The input unit can be used to receive input digital or character or image information, and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control. Specifically, the input unit of this embodiment includes not only a camera, but also a touch-sensitive surface (such as a touch display) and other input devices.
[0141] The display unit can be used to display information input by the user or information provided to the user and various graphical user interfaces of the device, which can be composed of graphics, text, icons, videos and any combination thereof. The display unit may include a display panel, and optionally, the display panel may be configured in the form of LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode), etc. Further, the touch-sensitive surface may cover the display panel, and when the touch-sensitive surface detects a touch operation on or near it, it is transmitted to the processor to determine the type of touch event, and then the processor provides a corresponding visual output on the display panel according to the type of touch event.
[0142] The embodiment of the present invention further provides a computer-readable storage medium, which may be a computer-readable storage medium included in the memory in the above embodiment; or a computer-readable storage medium that exists independently and is not installed in a device. The computer-readable storage medium stores at least one instruction, which is loaded and executed by a processor to implement Figure 1 The semantic encryption transmission method of the perception information and communication information of the synaesthesia integration platform shown in the figure. The computer readable storage medium can be a read-only memory, a disk or an optical disk, etc.
[0143] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the device embodiment, equipment embodiment and storage medium embodiment, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0144] Furthermore, in this document, the terms "comprises," "comprising," or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device that includes the element.
[0145] The above description shows and describes the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the invention, through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not depart from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.
Claims
1. A semantic encryption transmission method for perception information and communication information of a synaesthesia integrated platform, characterized in that: The method comprises a perception semantic encryption processing branch and a communication semantic encryption processing branch; wherein the perception semantic encryption processing branch comprises: Collecting and preprocessing the perception information to obtain perception candidate keywords, and selecting perception semantic keywords from the perception candidate keywords; Determining sensitive perceptual semantic keywords among the perceptual semantic keywords, encrypting the sensitive perceptual semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keywords by calculating the average channel confidentiality rate; further comprising: Determining sensitive perceptual semantic keywords among the perceptual semantic keywords according to the perceptual information sensitivity; Using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive perception semantic keywords; according to Calculate the average confidentiality rate of the channel, and use the channel with the highest average confidentiality rate as the secure transmission path to transmit the perceptual semantic keyword, where R represents the average confidentiality rate of the channel, the superscript + represents the positive part of the operation result, and η n {0,1} indicates whether the receiver receives data in the nth time slot, SNR n,u and SNR n,s They are used to calculate the equivalent signal-to-noise ratio of the receiver to the sensor and the equivalent signal-to-noise ratio between the listener and the sensor in the nth time slot. lb(˙) represents the logarithm operation with base 2, which is equivalent to log2(˙); Recovering and reconstructing the received perceptual semantic keywords using a preset perceptual semantic library, and outputting the recovered perceptual information; The communication semantic encryption processing branch includes: Collecting and preprocessing communication information to obtain candidate communication keywords, and selecting communication semantic keywords from the candidate communication keywords; Determining sensitive communication semantic keywords among the communication semantic keywords, encrypting the sensitive communication semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating an average channel confidentiality rate; further comprising: Determining sensitive communication semantic keywords among the communication semantic keywords according to the sensitivity of the communication information; Using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive communication semantic keywords; according to Calculating an average confidentiality rate of a channel, and using a channel with the highest average confidentiality rate of a channel as a secure transmission path to transmit the communication semantic keyword; Recovering and reconstructing the received communication semantic keywords using a preset communication semantic library, and outputting the obtained communication information; Wherein, determining the sensitive perceptual semantic keywords in the perceptual semantic keywords according to the perceptual information sensitivity specifically includes: dividing the perceptual semantic keywords into a p The keyword classification problem with sensitivity is represented by f p1 ,…,f pt F p , the keyword sensitivity target value is y p The goal of this classification problem is to select the k most sensitive p Keywords, define x p =(x p1 ,…,x pn ) {0,1} n Denotes the solution to the sensitive keyword selection problem, and defines Q p is an upper triangular matrix or n The symmetric matrix of n, Q p The expression of is calculated using the following formula: ; Among them, r(·,·) is the Pearson correlation coefficient. The support vector classifier is used to evaluate the sensitivity level. First, a support vector classifier is trained for each keyword sensitivity using the semantic keyword samples. Then, each trained classifier is used to classify each sensitive keyword into a sensitivity level to obtain the sensitivity result corresponding to the transmission-aware semantic keyword. Set up a quadratic unconstrained binary optimization problem, and add a quadratic penalty term as the cost function of sensitive semantic keyword selection to make the actual number of selected keywords deviate from k p When , the value of the cost function increases, and the resulting optimization problem based on sensitive perception semantic keyword selection is calculated using the following formula: , where represents the transposition operation of the solution to the sensitive keyword selection problem, Q p represents an upper triangular matrix or a symmetric matrix, x p represents the solution to the sensitive keyword selection problem, t p Indicates the number of keyword sensitivity levels, x pi represents the selected i-th sensitive keyword, k p Indicates the number of selected sensitive keywords; The cost function is estimated by executing the quantum approximate optimization circuit model q times, and the Hamiltonian expectation of the final state of the system after each execution is measured to evaluate the pros and cons of each solution. Finally, the solution with the lowest cost is selected, which represents the best choice for the most sensitive semantic keyword. The process of determining the sensitive communication semantic keywords among the communication semantic keywords according to the sensitivity of the communication information is the same as the above-mentioned process of determining the sensitive perception semantic keywords.
2. According to claim 1, the semantic encryption transmission method of perception information and communication information of the synaesthesia integration platform is characterized in that: The collecting and preprocessing of the perception information to obtain the perception candidate keywords, and selecting the perception semantic keywords from the perception candidate keywords, includes: Dividing the perception information into complete sentences and filtering each sentence to obtain the perception candidate keywords; The perception candidate keywords are defined as a directed graph G including nodes and edges p =(V p , E p ), where V p represents the perception candidate keyword node set composed of the perception candidate keywords, E p It is a V p V p A set of , which represents the association value between the perceived candidate keyword nodes; according to Calculate the score of each perceived candidate keyword node based on the directed graph, where is the damping coefficient, V pm Represents the point-to-point perception candidate keyword node V pi The node V pi 、V pj Represents the perceived candidate keyword node. For each perceived candidate keyword node V pi ,In(V pi ) indicates pointing to V pi The set of perceived candidate keyword nodes, Out(V pi ) indicates V pi The set of perceived candidate keyword nodes pointed to, , represents weight, WS(V pj ) indicates V pj The perceived candidate keywords of are based on the scores of the directed graph; The weights of the perception candidate keyword nodes are iteratively propagated according to the scores until convergence, and the weights of each perception candidate keyword node are sorted, and the N perception candidate keywords with the highest weights are selected as the perception semantic keywords.
3. The semantic encryption transmission method of perception information and communication information of a synaesthesia integration platform according to claim 1 is characterized in that: The method of restoring and reconstructing the received perceptual semantic keywords by using a preset perceptual semantic library and outputting the obtained restored perceptual information includes: Using a dynamic multi-layer semantic perceptron model to semantically classify the perceptual semantic keywords, and obtain a perceptual semantic segmentation result; Performing feature matching on each of the perceptual semantic keywords in the perceptual semantic segmentation result based on the perceptual semantic library, mapping the perceptual semantic keywords to the perceptual information space corresponding to the perceptual semantic library, and reconstructing through the features of the perceptual information space to obtain restored perceptual information; A first mutual information between the restored perception information and the perception information is calculated, and a similarity between the restored perception information and the perception information is evaluated according to the first mutual information.
4. The semantic encryption transmission method of perception information and communication information of a synaesthesia integration platform according to claim 1 is characterized in that: The collecting and preprocessing of communication information to obtain communication candidate keywords, and selecting communication semantic keywords from the communication candidate keywords, includes: Dividing the communication information into complete sentences and filtering each sentence to obtain the communication candidate keywords; The communication candidate keywords are defined as a directed graph G including nodes and edges c =(V c , E c ), where V c represents the communication candidate keyword node set composed of the communication candidate keywords, E c It is a V c V c A set of,represents the association value between the communication candidate keyword nodes; according to Calculate the score of each communication candidate keyword node based on the directed graph, where is the damping coefficient, V m Indicates the candidate keyword node V pointing to the communication ci The node V ci 、V cj Represents the communication candidate keyword node. For each communication candidate keyword node V ci ,In(V ci ) indicates pointing to V ci The set of perceived candidate keyword nodes, Out(V ci ) indicates V ci The set of perceived candidate keyword nodes pointed to, , represents weight, WS(V cj ) indicates V cj The correspondence candidate keywords of the directed graph are scored based on; The weight of each communication candidate keyword node is iteratively propagated according to the score until convergence, and the weight of each communication candidate keyword node is sorted, and the N communication candidate keywords with the highest weight are selected as the communication semantic keywords.
5. The semantic encryption transmission method of perception information and communication information of a synaesthesia integration platform according to claim 1 is characterized in that: The method of restoring and reconstructing the received communication semantic keywords by using a preset communication semantic library and outputting the obtained communication information includes: Using a dynamic multi-layer semantic perception model to semantically classify the communication semantic keywords, and obtain a communication semantic segmentation result; Performing feature matching on each of the communication semantic keywords in the communication semantic segmentation result based on the communication semantic library, and mapping the communication semantic keywords to the communication information space corresponding to the communication semantic library, and reconstructing through the features of the communication information space to obtain restored communication information; The second mutual information between the restored communication information and the communication information is calculated, and the similarity between the restored communication information and the communication information is evaluated according to the second mutual information.
6. A semantic encryption transmission device for perception information and communication information of a synaesthesia integrated platform, characterized in that: The device comprises a perception semantic encryption processing branch and a communication semantic encryption processing branch; wherein the perception semantic encryption processing branch comprises: A first preprocessing unit, configured to collect and preprocess the perception information to obtain perception candidate keywords, and select perception semantic keywords from the perception candidate keywords; A perceptual encryption unit, used for determining sensitive perceptual semantic keywords among the perceptual semantic keywords, encrypting the sensitive perceptual semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the perceptual semantic keywords by calculating the average channel confidentiality rate; further comprising: Determining sensitive perceptual semantic keywords among the perceptual semantic keywords according to the perceptual information sensitivity; Using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive perception semantic keywords; according to Calculate the average confidentiality rate of the channel, and use the channel with the highest average confidentiality rate as the secure transmission path to transmit the perceptual semantic keyword, where R represents the average confidentiality rate of the channel, the superscript + represents the positive part of the operation result, and η n {0,1} indicates whether the receiver receives data in the nth time slot, SNR n,u and SNR n,s They are used to calculate the equivalent signal-to-noise ratio of the receiver to the sensor and the equivalent signal-to-noise ratio between the listener and the sensor in the nth time slot. lb(˙) represents the logarithm operation with base 2, which is equivalent to log2(˙); A perception reconstruction unit, used to restore and reconstruct the received perception semantic keywords using a preset perception semantic library, and output the obtained restored perception information; The communication semantic encryption processing branch includes: A second preprocessing unit is used to collect and preprocess communication information to obtain candidate communication keywords, and select communication semantic keywords from the candidate communication keywords; A communication encryption unit, used for determining sensitive communication semantic keywords among the communication semantic keywords, encrypting the sensitive communication semantic keywords, and selecting a channel with the highest average channel confidentiality rate to transmit the communication semantic keywords by calculating an average channel confidentiality rate; further comprising: Determining sensitive communication semantic keywords among the communication semantic keywords according to the sensitivity of the communication information; Using a quantum approximate optimization algorithm to find an optimal encryption strategy to encrypt the sensitive communication semantic keywords; according to Calculating an average confidentiality rate of a channel, and using a channel with the highest average confidentiality rate of a channel as a secure transmission path to transmit the communication semantic keyword; A communication reconstruction unit, used to recover and reconstruct the received communication semantic keywords using a preset communication semantic library, and output the obtained communication information; Wherein, determining the sensitive perceptual semantic keywords in the perceptual semantic keywords according to the perceptual information sensitivity specifically includes: dividing the perceptual semantic keywords into a p The keyword classification problem with sensitivity is represented by f p1 ,…,f pt F p , the keyword sensitivity target value is y p The goal of this classification problem is to select the k most sensitive p Keywords, define x p =(x p1 ,…,x pn ) {0,1} n Denotes the solution to the sensitive keyword selection problem, and defines Q p is an upper triangular matrix or n The symmetric matrix of n, Q p The expression of is calculated using the following formula: ; Among them, r(·,·) is the Pearson correlation coefficient. The support vector classifier is used to evaluate the sensitivity level. First, a support vector classifier is trained for each keyword sensitivity using the semantic keyword samples. Then, each trained classifier is used to classify each sensitive keyword into a sensitivity level to obtain the sensitivity result corresponding to the transmission-aware semantic keyword. Set up a quadratic unconstrained binary optimization problem, and add a quadratic penalty term as the cost function of sensitive semantic keyword selection to make the actual number of selected keywords deviate from k p When , the value of the cost function increases, and the resulting optimization problem based on sensitive perception semantic keyword selection is calculated using the following formula: , where represents the transposition operation of the solution to the sensitive keyword selection problem, Q p represents an upper triangular matrix or a symmetric matrix, x p represents the solution to the sensitive keyword selection problem, t p Indicates the number of keyword sensitivity levels, x pi represents the selected i-th sensitive keyword, k p Indicates the number of selected sensitive keywords; The cost function is estimated by executing the quantum approximate optimization circuit model q times, and the Hamiltonian expectation of the final state of the system after each execution is measured to evaluate the pros and cons of each solution. Finally, the solution with the lowest cost is selected, which represents the best choice for the most sensitive semantic keyword. The process of determining the sensitive communication semantic keywords among the communication semantic keywords according to the sensitivity of the communication information is the same as the above-mentioned process of determining the sensitive perception semantic keywords.
7. A semantic encryption transmission device for perception information and communication information of a synaesthesia integrated platform, characterized in that: It includes a processor, a memory and a computer program stored in the memory, wherein the computer program is executed by the processor to implement the steps of a semantic encryption transmission method of perception information and communication information of a synaesthesia integration platform as described in any one of claims 1 to 5.
8. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by the processor, the steps of the method for semantically encrypting and transmitting perception information and communication information of a synaesthesia integration platform as described in any one of claims 1 to 5 are implemented.
